### 1. Introduction

^{1, 2)}. The bead geometry is directly influenced by the welding process parameters

^{2-5)}. To avoid weld bead defects and insure satisfactory mechanical properties, it is therefore necessary to carefully set-up the process parameters. These parameters are welding current, arc voltage, welding speed, torch angle, free wire length, nozzle-to-plate distance, welding direction, position and the flow rate and composition of the shielding gas

^{6)}.

^{1-3,6-12)}. The Response Surface Methodology (RSM)

^{13)}is one of those methods and has been widely used to study the weld bead geometry as a function of several process parameters

^{1,2,6-12)}.

### 2. Experimental section

### 2.1 Cladding process

^{-1}). The cladding material was an AWS classification E-7012 solid wire (OK Autrod 13.12) with a diameter of 1.2 mm. The base metal were 20×20×100 mm plates of C-CH35ACR low carbon steel. The chemical compositions of the base material and the cladding wire are given in Table 1.

##### Table 1

C | Si | Mn | Cr | Ni | Cu | Mo | P | S | |
---|---|---|---|---|---|---|---|---|---|

Base metal: C-CH35ACR mild steel | 0.01-0.03 | 0.08-0.1 | 0.2-0.4 | 0.1 | 0.1 | 0.1 | 0.03 | 0.04 | |

Welding wire: OK Autrod 13.12 | 0.1 | 0.5 | 1.1 | 0.5 | 0.5 | 0.2 | 0.03 | 0.03 |

^{4, 5)}. The wire feed rate and the welding arc current accordingly were chosen to obtain similar welding bead profiles according to different welding speed. The faster the welding speed, the fewer the melted wire deposited per unit length and therefore the smaller the bead geometry.

^{5, 14)}. Boiko at al.

^{14)}studied the effect of shielding gases on the MAG welding process. The different thermal conductivity of the shielding gases has a considerable influence on the arc configuration and the bead geometry. In the present study, different shielding gases were used with between argon, CO

_{2}, O

_{2}and mixture gases at a constant flow rate of 15 L/min. The standard names and composition of the gases are given in Table 2.

### 2.2 Experimental design

^{13)}and was similar to previous similar studies

^{2,7-11)}. This empirical method is commonly used for process in an industrial setting to optimize a response (here the bead geometry) influenced by several independent variables (here the process parameters). This method has been found to be valuable for the particular case of GMAW optimization

^{1)}. As described for example by Bezerra et al.

^{15)}, some stages in the application of RSM include: (1) selection of independent variables of major effects on the system and delimitation of the experimental region, (2) choice of the experimental design and experimental runs according to the selected design matrix, (3) mathematical-statistical treatment of the obtained experimental data through the fit of a polynomial function, (4) evaluation of the model’s fitness, (5) evaluation of the optimum values for each studied variables.

#### 2.2.1 Process variables and response

#### 2.2.2 Limits of the process variables

#### 2.2.3 Design matrix

^{4}(16) factorial design, 8 star points (one variable at its highest level +2 or lowest level -2 with all the other variables at the intermediate level 0) and 8 center points (all variables at the intermediate level 0). In this way, the 32 experimental runs allowed the estimation of linear, quadratic and linear-linear interactive effects of the welding parameters on the bead geometry. The design matrix is given in Table 4.

##### Table 4

#### 2.2.4 Experimental work according to the design matrix and record of the responses

^{16)}. The heat input is typically calculated as the ratio of the power to the velocity of the heat source, as described in Eq. (3):

*HI*= f

_{GMAW}=0.86 is the heat transfer efficiency

^{17)},

*I*is the welding current (A),

*U*is the arc voltage (V) and S is the welding speed (cm/min).

#### 2.2.5 Development of the mathematical models

*Y*is the response and

*S*,

*U*,

*I*and

*SG*are the welding speed, the arc voltage (U), the welding current (I) and the shielding gas (SG) respectively. A second-degree polynomial equation, a model commonly used in RSM

^{2,7,9,15,18)}, was selected for the four input variables to represent the response and is given in Eq. (5):

*Y*is the response, β

_{0}is the constant term, β

*represents the coefficient of the linear parameters, β*

_{i}*represents the coefficients of the quadratic parameter, β*

_{ii}*represents the coefficients of the interaction parameters,*

_{ij}*x*(or

_{i}*x*) represents the input variables (

_{j}*S*,

*U*,

*I*and

*SG*) and

*ε*is the residual associated to the experiments. The values of the coefficients in Eq. (5) were calculated by fitting the functions to the data using a statistical software (STATISTICA, Excel S-PLUS 8.0). A computer program was also developed to calculate the value of these coefficients for different responses (Table 5).

##### Table 5

##### (10)

##### (11)

##### (12)

### 3. Results and Discussion

### 3.1 Direct effects of the welding parameters on the weld bead

^{4, 5)}.

_{2}in the shielding gas increases. From Fig. 2A, Fig. 2 it was seen that the width decreases when the welding speed increases, as discussed earlier.

### 3.2 Interaction effects of the welding parameters on the weld bead

_{2}is lower in the mixture. The nature of the shielding gases also shows great influence on the contact angle when varying the welding speed (Fig. 3J), in particular when the welding speed is between 45 and 85 cm/min.

^{4, 5)}. The dilution should be as low as possible. As it was mentioned before, high dilution is suitable for weld joint while low dilution is preferred in cladding or surfacing. When the dilution is low the final composition of the deposited material is closer to that of the actual filler material. Ideally, the dilution in such cladding process should be lower than 50 %. In the present study, considering the results presented in Figs. 2C, 3C, 3D, and 3I, the upper limit for the dilution (D) should be no more than 57 %. The contact angle (θ) also has a significant influence on the mechanical properties of the weld. It is usually sought a high contact angle of about 70° to 90°, where the bead profile looks like half a disc. To insure satisfying conditions in the present study and considering Figs. 2A, 2D, 3E, 3F and 3J, the contact angle should be higher than 45°. At last, the heat input (HI) influences the cooling rate and in turn the microstructure of the bead. In the conditions of the present study, considering Figs. 2A, 2F and 3H, the heat input should be close to 400 J/mm or more. Based on these consideration and the results given graphically according to the mathematical model built in section 2.2.5, the following conclusions can be drawn regarding the welding parameters considered in this study: the welding speed should be S ≈ 60 cm/min, the welding current should be I ≈ 200-250 A, the arc voltage should be U ≈ 20-25 V and the shielding gas should be SG = M21 (80% Ar-20% CO

_{2}) for GMAW cladding on low-carbon steel.

### 4. Conclusion

_{2}).